February 21, 2012 at 3:50 pm
I want to know if it is possible to partition data in sql 2005 using the uinqueidentifier field. If yes how to go about it and will it have any performance issues?
Thanks
CP
February 22, 2012 at 9:40 am
Yes, it's possible (as per BOL: "Any column whose data type can be used as an index key can also be specified as a partitioning column, except the timestamp data type") But you should think twice before doing this, I mean how you create and use Partition function with uniqueidentifier data type
February 22, 2012 at 9:53 am
Be aware the uniqueidentiier values are random and therefore I would have thought not a good value to partition by.
If you have to use this use it in conjunction with NEWSEQUENTIALID ( )
so the values are increasing
---------------------------------------------------------------------
February 23, 2012 at 7:41 am
1) GUIDs have many VERY bad downsides for performance
2) Partitioning is a VERY advanced and complex topic. PLEASE do yourself and your company a favor and don't just read a few blog posts and BOL and slap something together and throw it into production. I can't tell you the number of clients/users I have come across that have messed things up and CAUSED problems by trying to do that. Note also that partitioning is NOT primarily a performance-enhancing mechanism. It exists mainly for maintenance/administrative reasons although it CAN provide a significant boost to throughput in some scenarios. Get a professional to help you evaluate your needs and come up with a plan and implementation and monitoring that will succeed!
Best,
Kevin G. Boles
SQL Server Consultant
SQL MVP 2007-2012
TheSQLGuru on googles mail service
Viewing 4 posts - 1 through 3 (of 3 total)
You must be logged in to reply to this topic. Login to reply